We were inspired by a gap in the world: we wanted something to conveniently track spending, but it was missing.
What it does
Scan your receipt with your camera, and have its details extracted to track your spending!
How we built it
Tesseract for OCRUsed in original version, replaced by GCV OpenCV for image processingUsed in original version, mooted by GCV
- Google Cloud Vision for OCR and image processing
- VSCode and WebStorm as editors
- React for the frontend
Challenges we ran into
- Poor quality of text recognition in anything but ideal conditions. Solution: Use aggressive text processing.
- Lack of knowledge of new image processing libraries. Solution: Quick on-the-fly learning.
- Library incompatibilities in npm. Solution: Choose different libraries.
Accomplishments that we're proud of
- We created this in 25 hours
- We seamlessly worked together as a team, having never met each other before
- We learned new frameworks from scratch in limited time
What we learned
- Optical character recognition (OCR)
- Image processing
- React custom events
What's next for Rite
- Expand spending history (e.g. exporting to spreadsheet)
Use Firebase ML for text recognitionUsed Google Cloud Vision instead
- And more...